Cover Extreme Events in the Atmosphere Basic concepts Academic year 2013-2014 ICTP Trieste - Italy Dario B. Giaiotti and Fulvio Stel 1
Outline of the lecture Definition of extreme weather event. It is not an easy task. Are extremes in frequency extremes in intensity too? Intensity Frequency probability density function. Milestones of extreme weather event definitions. Examples of extreme event definitions Extreme events and risks. 2
Definition of extreme weather event: an example Extreme means far from the bulk, at the border, or not common. There is no unique definition of extreme event, in particular for weather phenomena: to define an extreme, a statistical distribution or a sample distribution is required. An example from IPCC s Fifth Assessment Report (AR5 2013). IPCC WG1 Annex III - Glossary. Extreme weather event An extreme weather event is an event that is rare at a particular place and time of year. Definitions of rare vary, but an extreme weather event would normally be as rare as or rarer than the 10th or 90th percentile of the observed probability density function. By definition, the characteristics of what is called extreme weather may vary from place to place in an absolute sense.... 3
Ho to define an extreme weather event It is necessary to define a weather event. It is necessary to have a set of weather events. It is necessary to group the events in homogeneous sub sets. Possibly to give a rank to the sub sets (to define an order relation in the set of sub sets). Example 4
Extremes in intensity and in frequency In most of cases extreme means both rare and at the border of the intensity range. Anyway this is not the rule. The definition of the weather event is the key point. Intensity - Frequency relations: monotonic function - extreme in frequency and in intensity; negative linear correlation - the simplest relation; general relation for all the weather events; Intensity Frequency 5
Extremes in intensity and in frequency (cont'd) In general, given a weather event, it is necessary to consider extremes in frequency and in intensity. Extremes in frequency this is related to the events repetition. Weather manifestations do not occur once only. Extremes in intensity this is related to the events ranking. The same weather event presents variability in its features. The basic tools for an extreme weather event preliminary analysis are: a) the frequency intensity joint probability function; b) the frequency intensity relation; 6
Extremes in intensity and in frequency: joint probability function In general, given a weather event, it is necessary to consider extremes in frequency and in intensity. Number of occurrences in the 30 years set Example: 30 years of event observations: (1) evaluate the intensity class for each event; (2) compute the monthly frequency for each year; (3) fill in the frequency - intensity table. 2-4 months; 3 events every month; 285 events every month; never Intensity monthly Frequency A B C D E F G 0.0 0.2 0.4 0.6 0.8 1.0 0 1 0 5 8 10 3 0 3 2 21 15 20 5 0 20 68 8 10 0 0 0 54 32 9 8 0 0 0 45 14 10 12 0 0 285 36 0 12 0 0 0 7
Intensity Frequency PDF prob(f,f+df;i, I+dI) = f(f,i)df di f(f,i) is the PDF Probability to have an event with intensity between I and I+dI with a frequency between F and F+dF 8
Example of general PDF General case for many atmospheric extreme events: negative linear correlation 9
Two tails extreme events PDF real case Examples: consecutive third wet dry day; windy days 10
Summarize the PDF information Conditional distribution (i.e. yearly, decadal or climatic fluctuations for a intensity class) Marginal distribution (i.e. regardless fluctuations inside the intensity class) 11
IPCC extreme events definition marginal distribution Recall the example from IPCC s Fourth Assessment Report (AR4 2007). Extreme weather event An extreme weather event is an event that is rare at a particular place and time of year. Definitions of rare vary, but an extreme weather event would normally be as rare as or rarer than the 10th or 90th percentile of the observed probability density function. By definition, the characteristics of what is called extreme weather may vary from place to place in an absolute sense....... It considers frequency only (marginal distribution)... it is assumes that extremes in frequency are also extremes in intensity (monotonic relation). 12
Extreme event definition milestones The definition of extreme event depends on information available for its identification, namely observational data, measurements, theory and numerical models outputs. Generally, the definition is function of space and time: Events that are extreme for an area of the world could not be for another one. Events that are extreme at one time may be not in the future or in the past. f(f,i;x;t) is the PDF The characterization of the extreme event should take into account for spatial and time event duration, not only the event intensity. The definition is a function of the impacts that an extreme event has on human activities and/or the environment. 13
Extreme event definition: an example from air quality Particulate matter suspended in the air PM10 (d 10 μm) and PM2.5 (d 2.5 μm) The EC and national laws recognize PM10 and PM2.5 have impacts on human health. This comes from scientific evidences. EC... establishes the need to reduce pollution to levels which minimize harmful effects on human health, paying particular attention to sensitive populations, and the environment as a whole, to improve the monitoring and assessment of air quality including the deposition of pollutants and to provide information to the public. [1] The EC directive 50/2008 confirms the need to limit PM10 and PM2.5 concentration and it highlights the impacts on human health: [1] [1] The 2008/50/EC which is a directive of the European Parliament and of the Council of the European Union 14
Extreme event definition: an example from air quality Operational numerical forecast of PM10 (d 10 μm) concentration over a municipality area (ARPA FVG - CRMA) Hourly concentration Daily average concentration EC directive 50/2008 daily limit 15
Extreme event definition: an example from tropical cyclones Saffir/Simpson and Australian scale classify the intensity of the same weather event but with different parameters and units. They have been tuned for different areas of the world Saffir/Simpson scale U. S. National Weather Service Australian scale Australia's Bureau of Meteorology 16
Definition of extreme weather event: towards a standard A complete definition of and extreme event requires: an objective and unambiguous identification of the event; the definition of the event intensity, which is a function of physical features of the events and of the risks the event generates (stressor role); the intensity frequency pdf. That is the statistics of event occurrence for each class of intensity. For poor statistics or monotonic frequency-intensity relations, marginal distribution for intensity is sufficient. 17
Extreme events and risks An extreme event is very often associated with a risk and this rise our attention on it. Talking about risks lead us to the definition of risk and its evaluation. (Again) There is no unique risk definition; here is an example. EPA definition A risk is the chance of harmful effects to human health or to ecological systems resulting form the exposure to an environmental stressor. A stressor is any physical, chemical, or biological entity that can induce an adverse response. Stressors may adversely affect specific natural resources or entire ecosystems, including plants and animals, as well as the environment with which they interact. So it is important to identify the stressor. With reference to atmospheric phenomena the stressor in an environmental one. Extreme events are stressors in many situations 18
Predictions scales and natural hazards The most important fraction of natural hazards is due to meteorological extreme events. 19
Risk and stressors The risk and the stressor are very often tied by operational criteria that are defined with laws. There are international laws, Federal (USA) or Community laws (EU), national and regional administration laws. Scientists and technicians work to define and quantify the criteria adopted by politicians in writing the laws. Risk assessment is, to the highest extent possible, a scientific process. In general terms, risk depends on the following factors: how much a stressor is present in the environment (quantity of stressor per unit time); how much exposure an ecological receptor has with the stressor (time of exposure); the harmful intensity of the stressor (damage per unit time and quantity, quality of stressor). 20
The stressor and vulnerability information contribute to risk analysis Cyclones accounts for a large fractions of natural disasters 21
Risk assessment process The risk assessment process usually begins by collecting measurements that characterize the nature and extent of stressor in the environment, as well as information needed to predict how the stressor behave in the future. Models + Measurements and observations +Theory Based on this, the risk assessor evaluates the frequency and magnitude of human and ecological exposures to the stressor, both now and in the future. This evaluation of exposure is then combined with information on the the expected response to a given level of exposure, to predict the probability, nature, and magnitude of the adverse (health) effects that may occur. All risk assessments should be based on a very strong base knowledge (i.e., reliable and complete data on the nature and extent of stressor, the magnitude and frequency of human and ecological exposure, and the harmful intensity of the stressor). 22
The contribution of the atmospheric hazards to natural disasters Note the observational bias 23 UNEP
Natural disasters statistics updated to 2010 and ten years average Database and definitions CRED defines a disaster as: a situation or event which overwhelms local capacity, necessitating a request to a national or international level for external assistance; an unforeseen and often sudden event that causes great damage, destruction and human suffering For a disaster to be entered into the database, at least one of the following criteria must be fulfilled: 10 or more people reported killed; 100 or more people reported affected; declaration of a state of emergency; call for international assistance. Source: Annual Disaster Statistical Review 2010 CRED and UNISDR (http://www.cred.be/sites/default/files/adsr_2010.pdf) UNISDR United Nations International Strategy for Disaster Reduction http://www.unisdr.org/ 24 CRED Centre for Research on the Epidemiology of Disasters WHO collaborating center (http://www.cred.be/)
Natural disasters statistics updated to 2012 and ten years average Disaster subgroups definition Many are related to atmospheric phenomena Source: Annual Disaster Statistical Review 2010 CRED and UNISDR (http://www.cred.be/sites/default/files/adsr_2010.pdf) UNISDR United Nations International Strategy for Disaster Reduction http://www.unisdr.org/ 25 CRED Centre for Research on the Epidemiology of Disasters WHO collaborating center (http://www.cred.be/)
Natural disasters statistics updated to 2012 and ten years average Source: Annual Disaster Statistical Review 2012 CRED and UNISDR (http://reliefweb.int/sites/reliefweb.int/files/resources/adsr_2012.pdf) UNISDR United Nations International Strategy for Disaster Reduction http://www.unisdr.org/ 26 CRED Centre for Research on the Epidemiology of Disasters WHO collaborating center (http://www.cred.be/)
Natural disasters statistics updated to 2010 and ten years average Number of events by continent and class Source: Annual Disaster Statistical Review 2010 CRED and UNISDR (http://www.cred.be/sites/default/files/adsr_2010.pdf) UNISDR United Nations International Strategy for Disaster Reduction http://www.unisdr.org/ 27 CRED Centre for Research on the Epidemiology of Disasters WHO collaborating center (http://www.cred.be/)
Natural disasters statistics updated to 2010 and ten years average Number of victims by continent and class Source: Annual Disaster Statistical Review 2010 CRED and UNISDR (http://www.cred.be/sites/default/files/adsr_2010.pdf) UNISDR United Nations International Strategy for Disaster Reduction http://www.unisdr.org/ 28 CRED Centre for Research on the Epidemiology of Disasters WHO collaborating center (http://www.cred.be/)
Natural disasters statistics updated to 2010 and ten years average Amount of damages by continent and class Source: Annual Disaster Statistical Review 2010 CRED and UNISDR (http://www.cred.be/sites/default/files/adsr_2010.pdf) UNISDR United Nations International Strategy for Disaster Reduction http://www.unisdr.org/ 29 CRED Centre for Research on the Epidemiology of Disasters WHO collaborating center (http://www.cred.be/)
Natural disasters statistics updated to 2012 and ten years average Number of events and victims the trend Source: Annual Disaster Statistical Review 2012 CRED and UNISDR (http://reliefweb.int/sites/reliefweb.int/files/resources/adsr_2012.pdf) UNISDR United Nations International Strategy for Disaster Reduction http://www.unisdr.org/ 30 CRED Centre for Research on the Epidemiology of Disasters WHO collaborating center (http://www.cred.be/)
Predictions scales and natural hazards The increase of the exposure results in a trend in costs. 31
Natural disasters the weight of the atmospheric related events Hydro-meteorological disasters are the main source of casualties Hydrological events are responsible for the largest fraction of victims 2007 year $US Million The geophysical hazards account for a significant fraction of damages Source: Annual Disaster Statistical Review 2010 CRED and UNISDR (http://www.unisdr.org/files/2796_credannualstatisticalreview2007.pdf) UNISDR United Nations International Strategy for Disaster Reduction http://www.unisdr.org/ 32 CRED Centre for Research on the Epidemiology of Disasters WHO collaborating center (http://www.cred.be/)
Limits of risk assessment In practice, information is usually limited on one or more of the key data needed for risk assessment calculations: the magnitude and frequency of human and ecological exposure reliable and complete data on the nature and extent of stressor the harmful intensity of the stressor This means that risk assessors have to make estimates in computing risk, as a consequence... All risk evaluations are uncertain to some degree. A fundamental and not negligible part of all good risk assessments is a fair and open presentation of the uncertainties in the calculations and a judgment on the reliability of the resulting risk estimates. 33
Risk assessment and risk management: a step further The management of the risk follows the risk assessment. The risk assessment provides information on risk. The risk assessment is a scientific process only. The risk management takes actions to minimize the adverse response. The risk management actions are based on risk assessment data and several other information. Risk management is a function of many not scientific information and several scientific ones. 34
Risk management in a snapshot Risk management is a process which is function of: risk assessment; economic factors that inform the manager on the cost of risks and the benefits of reducing them, the costs of risk mitigation or remediation options and the distributional effects; laws and legal decisions are factors that define the basis for the risk assessments, management decisions, and, in some instances, the schedule, level or methods for risk reduction; social factors, such as income level, ethnic background, community values, land use, zoning, availability of health care, life style, and psychological condition of the affected populations, may affect the susceptibility of an individual or a definable group to risks from a particular stressor; technological factors include the feasibility, impacts, and range of risk management options; political factors are based on the interactions among branches of the Federal/Community government, state, and local government entities, and even with foreign governments. 35
Extreme events indexes Extreme events often are summarized in an index Indexes are important because they summarize the extreme event, in particular the complex ones, in a scalar. Indexes allow comparison among extremes and then produces a event ranking which is relevant relation for a set. Indexes allow to downscale large scale forecasts. The impacts-relevant indexes of the extremes are the base for the evaluation of risks from extreme event forecasts, both deterministic and probabilistic. In developing indexes it is important to keep in mind that indexes are regional specific, human or biological activity dependent. Indexes may be defined on measurements and observational data, but not necessarily only on that. 36